JP3626144B2 - Method and program for generating 2D image of cartoon expression from 3D object data - Google Patents

Description

【００２７】
式（１）の右辺の第１項は、Ｌａｍｂｅｒｔ則に従った、光源と面との傾きに対してその面の輝度がどのように依存するかを表現した項である。また、式（１）の右辺上段の第２項や、右辺下段は、光源からの直接光が無いときの、周囲の光の照度Ｌ_ａｍｂとその周囲の光に対する係数Ｍ_ａｍｂによる効果を表現する項である。Ｌａｍｂｅｒｔ則は、その面の位置における照明によるベクトルＬ（その面の位置から光源に向かっていて、その光源の光束がその面に与える照度を長さとするベクトル）と、オブジェクトの面の外側に向けて立てたその面の法線ベクトルＮとの内積（Ｎ・Ｌ）に、そのオブジェクトの面の拡散係数Ｍ_ｄｉｆを掛けたものである。拡散係数は、面の材質、色等によって定まる係数で、反射率と光沢の情報を含んでいる。なお、光源の色と面の色との関係は別途考慮することができる。例えば光源が赤い光のときと青い光のときを考え、それぞれの光源が同じ方向から同じ照度をオブジェクトのその面に与えるものであっても、その面が例えば赤い面のときには、赤い光に対する反射される反射光が、青い光に対する反射光よりも高い輝度になる。光源と面の色のこういった効果を考慮した場合には、より精密には、Ｒ、Ｇ、Ｂの三原色の各色ごとにこの式を評価する。また、測光学・測色学的な正確さをさらに反映させることも不可能ではない。また、より簡便な手法として、グローシェーディングやフォンシェーディングなどによって面の陰影を求めても良い。
【００２８】
ブロック２２のレンダリングでは、仮想スクリーン面の全ての（ｘ、ｙ）座標において、三原色Ｒ，Ｇ，Ｂのそれぞれについて画素値Ｃ_ｉ（ｘ、ｙ）（ただし、ｉはＲ，Ｇ、Ｂ）が計算される。画素値Ｃ_ｉは典型的には０〜２５５まで（８ビット値をとる場合）であり、必要に応じてビット幅は増減される。これと同時に、オブジェクトが描かれたか描かれていないかを示すマスクＭ（ｘ、ｙ）が計算される。マスク（マスク情報）Ｍは、式（２）にしたがって算出される。マスクＭはその画素について面のレンダリングが行われれば１、行われなければ０となるデータであり、後にオブジェクトの輪郭を描画する際に用いられる。
【数２】

【００２９】
ブロック２２を実行する装置では、２次元画像フレームに対応して画素値用メモリとして働く任意の記憶装置が備えられており、この記憶装置に対して演算装置がレンダリングしたデータであるＣ_ｉ（ｘ、ｙ）を順次格納させてゆく。また、同様に全画素用のアドレス空間を持つマスク用メモリに対してＭ（ｘ、ｙ）が格納される。
【００３０】
その後、後に詳細に説明する主線・輪郭描画ステップ（ブロック２３）によって主線と輪郭が求められる（ブロック２３）。この処理を経ることによって、仮想空間内の数値表現であるオブジェクトから図６（ａ）および図７（ａ）のような主線（輪郭を含む）で表現される２次元画像データが得られる。図６（ａ）の線６１は球オブジェクトから生成された輪郭線で、図７（ａ）の線７１は多面体オブジェクトから生成された輪郭線、線７６は主線である。このような主線・輪郭描画はオブジェクトがより複雑なものであっても同様である。
【００３１】
また、後に詳細に説明する陰影処理ステップ（視覚的階調化ステップ、ブロック２４）によって、連続階調の陰影を１以上のテクスチャで表現される。テクスチャは、例えば、図６（ｂ）や図６（ｃ）に示すドットのパターンや、図示しないストライプといった、従来のスクリーントーンにみられるようなもの、あるいは、一定の色相や明度を持つ「ベタ塗り」などである。これにより、図６（ｂ）や図６（ｃ）に示したように、一定の平均明度を有する少なくとも１つ以上のテクスチャで陰影を表現する。
【００３２】
ブロック２３、ブロック２４によって主線や輪郭が作られた画像（図６（ａ）、図７（ａ））と、陰影がテクスチャによって表現された画像（図６（ｂ）、（ｃ）、図７（ｂ）、（ｃ））とは、画像表示ステップ（ブロック２５）で合成されて全体として漫画的表現を含んだ画像が得られる。この例を図６（ｄ）、図７（ｄ）に示している。
【００３３】
以上から分るように、図６（ａ）や図７（ａ）のみでは単なる主線や輪郭線のみを示すものであるし、主線を用いずに図６（ｂ）と図６（ｃ）とを組み合わせただけのものや、あるいは、図７（ｂ）と図７（ｃ）とを組み合わせただけのものでは、面の明るさがテクスチャで示されるのだけなので、複雑なオブジェクトの場合に形状を把握することは難しい。また、図示していないが、レンダリング直後の連続的なリアルな陰影を持った画像に、例えば図６（ｂ）と図６（ｃ）をただ単に合成しても、リアルな画像にスクリーントーンを張っただけのようになり、漫画的表現とは程遠い。
【００３４】
［主線・輪郭描画］
本発明の特徴である主線（輪郭線を含む）を描画するステップ（ブロック２３）について詳細に説明する。この主線・輪郭描画ステップを図３に説明する。このステップでは、まず、２次元画像フレーム内での注目画素を指定し（ブロック３０）、オブジェクトの面の法線ベクトルを抽出する（ブロック３１）。例えば立体オブジェクトがポリゴンで表現されているときには、オブジェクトの外形面を表す少なくとも３つの頂点Ａ，Ｂ，Ｃを持つ面５（図５）では、その頂点のそれぞれに対して、法線ベクトルＮが定まる。この法線ベクトルＮを頂点以外の点（例えば図５のＤ）は、適当な数学的な方法をもちいて、頂点Ａ，Ｂ，Ｃの法線ベクトルの値から補間して求めることができる。これによりオブジェクトの任意の面の任意の点に対して法線ベクトルが定まる。この法線ベクトルは、仮想空間においてＸ，Ｙ，Ｚの各方向に対応して（Ｎ_Ｘ，Ｎ_Ｙ，Ｎ_Ｚ）という成分で表現される。法線ベクトルを抽出する必要があるのは、最終的に見える面だけであり、視点からは隠れていて陰面処理されてしまう面に対しては算出の必要がない。このようにして、２次元画像フレーム内の位置（ｘ、ｙ）に対して、その位置での外形面の法線ベクトルＮを帰属させることができる。つまり、ｊをＸ，Ｙ，Ｚとして、Ｎ_ｊ（ｘ、ｙ）を求めることができる。これを法線ベクトル用メモリ等の適当な記憶装置に格納する。オブジェクトの面が他の方法によって定まっているときには、適宜他の方法によって法線ベクトルを求めればよい。
【００３５】
また、ブロック３２で奥行き情報（距離情報）を抽出する。奥行き情報とは、仮想空間における仮想スクリーンからその画素に表示されるオブジェクトの外形面までの距離か、あるいは、視点からその外形までの距離である。視点とオブジェクトとを結ぶ視線に沿って求める。レンダリング処理における陰面処理等において用いられるｚバッファの値があればその値を用いても良い。法線ベクトルとは異なり、スカラー量であるが、２次元画像フレーム内の各画素（ｘ、ｙ）に対して、ｚ（ｘ、ｙ）として求まる。これを奥行き情報用メモリに蓄えてもよいし、ｚバッファの値をそのまま用いても良い。
【００３６】
以上のようにして、２次元画像フレーム内の各画素（ｘ、ｙ）に対して、マスクＭ，画素値Ｃ_ｉ、奥行きｚ、法線ベクトルの各成分Ｎ_ｊが求まる。
【００３７】
主線は、２次元画像フレームにおいて隣り合った面同士が仮想空間で不連続な場合と、連続であるが滑らかでない状態でつながる場合とにおいて、その境界部分に描画する。また、輪郭部分にも描画する。図７（ａ）の多面体オブジェクト７０における面７８と面７９が２次元画像フレーム１００において隣りあっているが、仮想空間では不連続である。また、面７７と面７９が２次元画像フレームにおいて隣りあっていて仮想空間で連続ではあるが、折れ曲がっていて滑らかには連続してない。漫画などでは従来からこのような面の境界は、制作者によって主線と呼ばれる線によって描かれてきた。
【００３８】
この主線を描くには、２次元画像フレーム内の各画素（ｘ、ｙ）に対して、その点の近傍とｚやＮ_ｊが変化していないかどうかの差分を求める（ブロック３３，３４）。これらは、ラプラシアンフィルタによって実行でき、ｚについては式（３）、Ｎ_ｊについては式（４）によって差分をとることができる。
【数３】

【数４】

【００３９】
この差分値を用いることにより、主線の有無とその線幅が求められる。例えば、式（５）によって、ｚから求まるＷ_ｚを、式（６）によって、Ｎから求まるＷ_Ｎをそれぞれ求め、それらの値に適当な関数を用いて線幅を求める。
【数５】

【数６】

【００４０】
ここで、［定数ｚ］、［定数Ｎ］は、一定以上の差分値が得られない場合には主線を描かないことを示すオフセット値である。Ｗ_ｚ、Ｗ_Ｎが負の値の時には線を描画しない。また、ブロック３５において、既に式（２）によって計算しておいたマスク情報Ｍ（ｘ、ｙ）を抽出する。
【００４１】
また、輪郭線を求める。マスクＭが求まっているのでこれを用いて輪郭が判定可能である（ブロック３６）。つまり、注目画素と周囲８画素を含めた９画素のマスク情報Ｍの値の和を計算し、０または９のときはその画素はオブジェクトの描画領域の内側か外側かであり、また、それ以外のときには、少なくとも注目画素か周囲の画素の何れかに、オブジェクトの２次元画像フレーム上への射影図形の境界が含まれていることを示している。従って、その注目画素を輪郭と判定することができる（ブロック３６）。ここで、輪郭係数Ｂという変数を用いる。この輪郭係数Ｂは注目画素が輪郭であると判定された場合に１より大きな値（オペレータが設定できる値）となり、輪郭以外では１となる。これにより主線の線幅を与える式と輪郭線の線幅を与える式をまとめて定義できて、線幅Ｗ（ｘ，ｙ）は式（７）のようになる。
【数７】

【００４２】
このようにして得られた線幅Ｗを用い、この線幅Ｗだけの半径を有する円を、２次元画像フレーム内に重ねられる適当なレイヤー（以後「主線レイヤー」）に、位置（ｘ、ｙ）を中心に描画する（ブロック３７）。２次元画像フレーム内で注目画素をスキャンすると、結果として主線や輪郭線となる。主線レイヤーは例えば主線用メモリ等の適当な記憶装置に格納される。最終的には、この主線レイヤーのデータを２次元画像フレームに表示等する。主線レイヤーは、適当に色を変更したりすることが可能であり、黒やその他の適当な色で描くことができる。また、漫画等の別途印刷する場合には、印刷段階でこの部分の色を選択してもよい。
【００４３】
このようにしてある注目画素に対して処理を終えた後、注目画素を移動させて（ブロック３８）、全画素にわたってスキャンして終了する（ブロック３９）。注目画素を隣接した画素に移動する際には、奥行き用、法線ベクトル用のメモリにアクセスする回数を低減させるように、例えば注目画素の直ぐ右隣に移動するのあれば、新しく周囲画素となる右側の３画素のみメモリから読み出して、注目画素と周囲画素の残りの値は、直前の位置の注目画素の際の値を利用することもできる。
【００４４】
また、上記の説明で記載した順序以外に、例えば、ブロック３１の直後にブロック３３を処理し、その後ブロック３２、ブロック３４と処理したりすることも可能である。ブロック３７の主線・輪郭描画までに必要なマスクＭ，奥行きの差分値ｄｚ、法線ベクトルの各成分の差分値ｄＮ_ｊの各値がその注目画素について求まっていれば良い。
以上のような主線・輪郭描画ステップによって、図６（ａ）および図７（ａ）が得られる。図７（ａ）には輪郭と主線が共に描かれているが、輪郭部分は主線より太く描かれている。
【００４５】
また、図示しないが、オブジェクトの種類によっては、オブジェクトを作る個々の面を囲んでそれを構成する境界線（面の縁取り）に、テクスチャマッピング等の手法で予め主線を描画しておくことができる。これは、オブジェクトが比較的少ない数の面で構成されていて直線の多面体の組み合わせになっているような場合（建物など）にはレンダリング段階に既存のテクスチャマッピング手法によって主線が表現されるために有効である。この場合、２次元画像フレームにおいて主線を明示することができるが、主線と輪郭線の判定ができず、また、湾曲した面では境界線以外が輪郭となる場合もある。この場合には、輪郭描画ステップをあわせて用いることで、輪郭を主線で表現できる。
【００４６】
［陰影の処理（視覚的階調化ステップ）］
また、本発明のもう１つの特徴である、陰影処理ステップ（視覚的階調化ステップ、ブロック２４）を図４に詳細に説明する。従来の漫画的表現では、絵に表れる影などは必ずしも陰影を忠実に連続階調で表現したりせずに、一以上のテクスチャで表現することが行われてきた。そのテクスチャとして、多くの場合は周期的な白黒等の２値の明暗を有するスクリーントーンと呼ばれるパターンが多用されてきた。そのスクリーントーンは、例えば円形ドットの網点の正方格子や三角格子状のものの場合（図６（ｂ）、図７（ｂ）はこれに類似のパターンを用いて描かれている）がある。これらは、スクリーン線数（１インチ＝２．５４ｃｍ当たりの線数）で決まる一定のピッチをもった黒等の着色された円が透明なシートに多数配列されて作られている。多くの場合、スクリーン線数６０から７０本程度のパターンが用いられる。この他に、ストライプやその他の様々なスクリーントーンが漫画的表現に用いられている。また、テクスチャとしては、一様なベタ塗りなども利用できる。本発明の陰影処理ステップ（視覚的階調化ステップ）は、限定しないが、こういったテクスチャをコンピュータ上で再現し、オブジェクトがリアルに表現している陰影を、それを用いて漫画のような表現のテクスチャを採用して表現するための処理である。
【００４７】
この陰影処理では、まず、注目画素が指定されて（ブロック４０）、必要に応じてグレースケール化処理（ブロック４１）が行われる。これは、例えば、式（８）に従って、２次元画像フレームの各点において、ビデオ信号のＹ信号を算出する式である。
【数８】

【００４８】
必要に応じて、ガンマ処理等をおこなって画像の明るさや階調性を調整する（不図示）。そして、その画素のＣ_Ｙ（ｘ、ｙ）を、予めオペレータが定めておたり、あるいは既定値として持っている基準値Ｔ_１、Ｔ_２等と比較する。基準値は、陰影を表現するテクスチャを収納するレイヤー（以後トーンレイヤーと呼ぶ）の数を定め、その数から１を除いた個数だけ指定する。
【００４９】
次に各基準値Ｔ_ｋ（ここに、ｋ＝１、２、・・・（トーンレイヤー数）―１）とＣ_Ｙとを比較する。その結果、例えば、トーンレイヤー数が３で、基準値がＴ_１＜Ｔ_２なる二つの値だけ定められている場合には、Ｔ_１、Ｔ_２を境界にして、その位置（ｘ、ｙ）に描画するトーンレイヤーを決める。例えば、Ｃ_Ｙ＜Ｔ_１のときには（ブロック４２）、注目画素をトーンレイヤー１に格納されたテクスチャで描画し（ブロック４３）、Ｔ_１＜Ｃ_Ｙ＜Ｔ_２のときは（ブロック４４）トーンレイヤー２に格納されたテクスチャで（ブロック４５）、Ｔ_２＜Ｃ_Ｙのときには（ブロック４６）、トーンレイヤー３に格納されたテクスチャで描画する（ブロック４７）。そして注目画素を移動させて（ブロック４８）、全画素をスキャンすることによって（ブロック４９）、２次元画像フレーム内にこの処理を施す。これによって、予めトーンレイヤー１、２、３のそれぞれに、円パターンの占有割合で定義される濃度を用いて、例えば、濃度５０％（暗い）、濃度２０％（やや暗い）、濃度０％（何もない）といったテクスチャを格納しておけば、階調を減らしてから各階調をスクリーントーンのようなテクスチャで表現するという漫画的表現に見られる陰影表現が可能となる。
【００５０】
図６（ｂ）、図７（ｂ）は、共に濃度２０％のスクリーントーンの表現を用いており、トーンレイヤー２にあるスクリーントーンを模したデータを描画する部分に書いたそのデータである。また、図６（ｃ）、図７（ｃ）は、共に濃度５０％のスクリーントーンの表現を用いたトーンレイヤー１から描画される同様のデータである。濃度０％のスクリーントーンの表現は、何も描かないのと同じであるため、特には示さない。なお、このテクスチャは、予め用意されて適当な記憶装置に準備されても良いが、何らかの数学的な関係を用いて適宜生成して、数値的に生成しても良い。例えば、判定値によって分けられたＣ_Ｙ値が同じところは同じ濃度の網点を生成させたりすることができる。あるいは、元の画素値Ｃ_ｉから色相を保って明度を落とすといった処理をして生成するテクスチャでも良い。
【００５１】
このようにしてブロック２３で準備された主線レイヤーとブロック２４で準備された少なくとも１以上のトーンレイヤーを、合成して画像として表示すれば（ブロック２５）図６（ｄ）、図７（ｄ）が得られる。これによってオペレータは画像として立体オブジェクトデータから生成された漫画的表現が施された２次元画像を確認することが可能となる。
【００５２】
この画像は、画像表示によって目的が達成されるような用途（例えばコンピュータゲームなど）では、必要に応じてオブジェクトをジオメトリエンジンで駆動しつつ、ブロック２２からブロック２５までの処理を行い、動画の個々のフレームを進めて連続した動画等を生成させる。
【００５３】
また、アニメーションの制作現場などでは、必要に応じてブロック２６でオペレータの要求するものかどうかが判定される。その結果によってブロック２２から２６が適当なＡ、Ｂ，Ｃのポイントから再実行されて、コンピュータ上で漫画的表現を含んだＣＧの処理作業が効率的に行われる。
【００５４】
また、最終的に画像でユーザーに提供されるものではなくて印刷して提供されるような用途（ＤＴＰ，漫画制作現場等）では、更に、印刷ステップ（ブロック２７）を含めて、その結果によってブロック２２から２７が適当なＡ、Ｂ，Ｃのポイントから再実行される。
【００５５】
以上に説明した本発明の方法において、主線・輪郭描画ステップ（ブロック２３）と陰影処理ステップ（ブロック２４）は、この順に行う必要はない。逆に陰影処理ステップを行った後に主線・輪郭描画ステップを行うこともできる。大抵の主線や輪郭線は、陰影のテクスチャに上書きされるので、画像表示用の２次元画像フレームメモリー（フレームバッファ）に直接書き込むこともできる。また、ブロック２３とブロック２４の内部で独立して注目画素をスキャンさせているが、差分処理で注目画素の周囲の画素のデータを用いることなどを調整した上で、スキャンするループを共通にしてもよい。このようにすれば、ｚの差分値、法線ベクトルの成分Ｎ_ｊの差分値等を２次元画像フレーム分だけ格納するメモリを節約することができる。
【００５６】
［実施例］
図８〜１１に示すように、建物のオブジェクトから漫画的表現を構成した。図８には、この建物の２次元画像フレーム内における配置が描かれている。なお、図６（ａ）、図７（ａ）と同様に、仮想空間内のオブジェクト自体は数値データであるが、説明のために本発明の主線・輪郭処理後の図形を示す。
このオブジェクトである建物８０は、本願発明の方法に従って仮想空間で配置している。また、このオブジェクトには、地面に見える部分８１も含まれている。
【００５７】
さらに、オペレータがそのオブジェクト８０に当てる照明を決定し、仮想空間でそのオブジェクトを観察する視点や画角を決定した。本実施例では、ほぼ真上においた平行光源を設定した。
この光源によって建物のオブジェクトでレンダリング（変換ステップ、画素値算出ステップ）を行った。レンダリング直後（図２のブロック２２の処理直後）の段階では、図示しないが、面ごとに大きく明るさが異なると共に、面の中でも場所によってその明るさが異なっていた（これは、図１１における屋根の部分に明るい部分８７と暗い部分８６となって表れている）。レンダリング直後の段階では、この明るさの違いは、連続した陰影（グラデーション）に感じられるようなものであり、いわゆるリアルな表現ではあるが、漫画等に見られる画像表現ではない。
【００５８】
このレンダリングの後、本発明の主線・輪郭の描画を行い、陰影の処理（視覚的階調化）を行った。主線・輪郭の描画は、本願の方法に従って、立体オブジェクトの各面間での距離情報と法線ベクトル方向とから、自動的に生成された。式（５）や式（６）の関数のパラメータ（係数、定数）を変更すれば、線の太さを変更できたり、主線として描画する境界を全て拾うように選んだり、あまり拾わないように選んだりすることが可能であった。図８において、線８２は本実施の形態で得られた輪郭線を表し、線８３は主線である。輪郭線係数Ｂは、主線が輪郭のとき１．５倍されるように設定した。
【００５９】
また、陰影は、空白も含めて３種類のテクスチャで表現した。図９はこの実施例で中間の明度の部分に割り当てられた３０％の網点パターンの配置８４でありレイヤー２に格納したテクスチャを必要な範囲に切り取ったものである。また、図１０は、暗い明度の部分に割り当てられた５０％の網点パターンの配置８５でありレイヤー１に格納したテクスチャを必要な範囲に切り取ったものである。レイヤー３はここでは空白としている。この割り当ては、ＲＧＢの画素値から求めたＣ_Ｙ（Ｙ値）の値が０〜３１までは５０％網点、３２〜１２７までは３０％網点、１２８〜２５５までは空白となるように、Ｔ_１を３２、Ｔ_２を１２８として得られたものである。各画素値は、オブジェクトの属性として決定していた面ごとの拡散係数（反射率と光沢情報を含む）、照明条件、照明と面との方向の関係、オブジェクトの作る影等によってレンダリングによって得られた画素値である。図９、図１０はともに、本発明の陰影の処理（視覚的階調化）を行って立体オブジェクトデータから自動的に生成された。
【００６０】
以上のようにして生成された主線・輪郭の画像とテクスチャの画像を合成して、図１１に示す漫画的表現された２次元画像が生成され、表示装置に表示された。図１１の部分８７と部分８６は、共に同じオブジェクトの同じ面であるが、レンダリング直後のこの部分に見られたグラデーションを反映して、異なるテクスチャで描画された。主線の描画は、式（５），（６）のパラメータを変更して、どの程度の細かいディテールまで主線として描画するか、容易に設定できた。また、輪郭係数を変更して、容易に輪郭に描かれる主線（輪郭線）の太さを協調することができた。
【００６１】
また、テクスチャを採用する基準値に用いる基準値Ｔ_１、Ｔ_２をオペレータが調整することによって、それまで網点のテクスチャで置き換えられる部分が変化し、容易に処理状態を調整することができた。テクスチャは、網点以外にも用意されていて、それを切り換えることで異なる印象の様々な陰影の表現が容易に確認できた。
【００６２】
また、オペレータは仮想空間でのオブジェクトの向きや配置を容易に変更でき、視点を変更することも容易であった。これにより、オブジェクトを再配置したり、パース処理（奥行き感）が容易に修正できた。また、同じオブジェクトを別の場面で向きを変更して用いたり、再利用することもできた。
以上のように、本願の特徴によって、立体オブジェクトデータからオペレータの意図する漫画処理された２次元画像が容易に生成できた。
【００６３】
【発明の効果】
本願の発明によれば、コンピュータグラフィクスの技術において、漫画等で従来から用いられている主線や輪郭を描画し、かつ、陰影には任意のパターンや一定階調のベタ塗りの色といったテクスチャーを用いる表現が可能になる。これにより、コンピュータグラフィクスで従来の表現を用いることができる。また、オブジェクトを容易に再配置できたりするようなコンピュータの利点が、漫画制作の現場やアニメーション製作現場で利用できる。
【図面の簡単な説明】
【図１】本発明の立体オブジェクトの配置方法を示すフローチャートである。
【図２】本発明の立体オブジェクトデータから２次元画像を生成する方法を示すフローチャートである。
【図３】本発明の主線・輪郭描画ステップの詳細を説明するフローチャートである。
【図４】本発明の陰影処理ステップの詳細を説明するフローチャートである。
【図５】本発明の法線ベクトルの算出の概念を示す斜視図である。
【図６】本発明の方法によって球のオブジェクトから漫画的表現の球の２次元画像が生成される様子を示す説明図である。
【図７】本発明の方法によって多面体のオブジェクトから漫画的表現の球の２次元画像が生成される様子を示す説明図である。
【図８】本発明の実施の形態に係る、建物のオブジェクトから生成された主線および輪郭から表現された２次元画像示す説明図である。
【図９】本発明の実施の形態に係る、建物のオブジェクトから生成されたテクスチャ（２０％網点）のパターンとそれで表現される部分を示す説明図である。
【図１０】本発明の実施の形態に係る、建物のオブジェクトから生成されたテクスチャ（５０％網点）のパターンとそれで表現される部分を示す説明図である。
【図１１】本発明の実施の形態に係る、建物のオブジェクトから生成された主線および輪郭と、テクスチャのパターンとから合成して生成される漫画的表現された建物の２次元画像を説明する説明図である。
【符号の説明】
２３ 主線・輪郭描画
２４ 陰影処理（視覚的階調化ステップ）
１００ ２次元画像フレーム
６０ 球オブジェクト
６１ 主線（輪郭）
７０ 多面体オブジェクト
８０ 建物オブジェクト
６０１、６０２、７０１、７０２、８４、８５ テクスチャ（網点パターン）[0001]
[Technical field to which the invention belongs]
The present invention relates to the field of CG (computer graphics), the field of DTP (desktop publishing), the field of animation production and the production of comics and comics, and the field of computer games. In particular, the present invention relates to a method for generating a two-dimensional image of a cartoon expression from data relating to a three-dimensional object in a virtual space, and a program therefor.
[0002]
[Prior art]
CG technology has been used in the production of animations and comics, comics, etc. due to the high performance of computers, lower prices, and the generalization of related software. Conventionally, in the CG field, a three-dimensional object (or object) has been expressed in a virtual space. In particular, the fact that an object can be created in a virtual space and the image can be checked on a computer while making various movements and changing the orientation and arrangement is expressed in two dimensions in the end, such as manga production. There is a great advantage that the production process is shortened even in the field of producing still images. In the field of computer games, a CG image is provided to a user by finally displaying the two-dimensional image on a display screen.
[0003]
In these fields, in addition to the desire to express the two-dimensional images seen by users more naturally, more precisely, and in real life, the expression techniques that have been used and are used in conventional comics and animations ( There is a desire to realize an image expression that makes use of (hereinafter referred to as “cartoon expression”).
[0004]
This cartoon expression is a method of expressing a line that is the boundary between multiple faces of a solid with a line (main line) that emphasizes it with a line of a certain thickness, among which a main line that expresses the boundary between the background and the solid in particular A method of emphasizing the image as an outline, or a method of expressing a shadow created by the position of a light source, a curved surface of a three-dimensional shape or the presence of another three-dimensional object with a screen tone, or expressing that part with a constant color tone with low brightness It is. In the past, the producers expressed these by combining them using manual work or DTP techniques as necessary, realizing a comic expression.
[0005]
[Problems to be solved by the invention]
However, there is no disclosure of a method for automatically generating this cartoon expression from the three-dimensional object data in the virtual space.
Japanese Patent Application Laid-Open No. 7-129762 discloses a technique for extracting and reconstructing contours and shadows from an image already expressed in two dimensions. In this method, a contour or shadow is added based on information of only the two-dimensional image. For this reason, even if applied to a CG image, the main line cannot be expressed well. For example, the main line cannot represent a boundary between two parallel surfaces that are different in distance (depth) from the viewpoint and do not change in brightness. This is because such a boundary between the two surfaces hardly gives contrast in a real photograph or a CG image. When a conventional computer is not used, the author has finished a cartoon expression by writing a main line as necessary on these boundaries, but such an expression cannot be realized by a computer.
[0006]
Japanese Patent Laid-Open No. 2001-319244 attempts to add a screen tone image to a three-dimensional object. According to this method, it is possible to use a screen tone expression for the object. However, since a main line and an outline are not generated from a three-dimensional object, a cartoon expression cannot be obtained simply by using a screen tone.
[0007]
In view of the above problems, the present invention provides a method for realizing the above-described expression method used in conventional comics and animation with an image generated from solid object data of a computer, and a program for performing the method It is.
[0008]
[Means for Solving the Problems]
In the present invention, an arrangement input step for receiving the arrangement of a three-dimensional object in the virtual space by the input means, a viewpoint setting step for receiving the position of the viewpoint in the virtual space by the input means, and inputting the characteristics of the illumination light source in the virtual space A light source setting step accepted by the means, a step of obtaining the direction in which the three-dimensional object exists from the viewpoint as mask information by the computing means, a step of obtaining the color indicated by the three-dimensional object from the viewpoint by the illumination light source as color information by the computing means, A step of obtaining a distance between the outer shape surface seen from the viewpoint of the three-dimensional object and the viewpoint as distance information by the computing means, a step of obtaining a normal vector in the virtual space of the outer surface as normal vector information by the computing means, and mask information And distance information and normal vector A main line drawing step for drawing a main line in a pixel on the two-dimensional surface by the calculation means based on the color information, and a calculation means for calculating different textures before and after the reference value on the two-dimensional surface based on the color information and a certain reference value. 2 of the cartoon expression from the three-dimensional object data comprising the visual gradation step employed by the computer and the image display step of obtaining a composite image of the main line on the two-dimensional surface and the texture employed and displaying the composite image. A method for generating a dimensional image and a program for realizing the method are provided.
[0009]
As a result, it is possible to improve the efficiency of the production process by the computer while realizing the cartoon expression in the CG image and realizing the conventional expression method in the production of the cartoon and animation.
[0010]
Here, the main line drawing step and the visual gradation step may not be in the order as described above. There may be a main line drawing step after the distance information, normal vector information, and mask information are obtained, and a visual gradation step after the color information is obtained. That is, in addition to the order described above, there may be a main line drawing step after the visual gradation step, or parallel processing may be performed using a parallel processing method of an appropriate processing system.
[0011]
In addition to the characteristics of the illumination light source itself, such as the color temperature, direction, illuminance, and point light source or parallel light source, the illumination light source has a geometrical characteristic such as the position and orientation of the illumination light source in the virtual space. Includes placement features.
[0012]
In the present invention, in the main line drawing step, the main line indicating the outline of the three-dimensional object has a line width different from the main line indicating other than the outline.
This makes it possible to express the outline of the object more clearly, making it easier to recognize the boundary between the object and the part placed behind it, and to more faithfully represent the traditional expression methods used in manga and animation production. realizable.
[0013]
Further, in the present invention, the main line drawing step includes a step of obtaining a difference value of distance information from the distance information between the pixel of interest on the two-dimensional surface and the surrounding pixels, and the pixel of interest on the two-dimensional surface and the periphery thereof. And calculating a difference value of the normal vector information from the normal vector information with respect to the pixel of the main line by a calculation means, and the main line drawing step is based on the difference value of the distance information and the difference value of the normal vector information. Drawing is determined by the calculation means.
[0014]
As a result, even if the 3D object data is defined by being surrounded by a certain surface, the main line used in comics, etc., can be obtained from the geometric shape of the 3D object, which has been used in manga and animation production. The expression method can be realized more faithfully.
[0015]
In addition, in the present invention, the visual gradation step performs gradation using a numerical value generation texture generated by the calculation means using the numerical value calculated from the color information and the reference value.
Thereby, the expression by the shadow texture can be easily changed by changing the reference value by some input means.
[0016]
In the present invention, the hardware environment in which the method and program shown here are realized is not particularly limited. Any device that has an arithmetic function, can handle images, can input and output files as necessary, has an appropriate storage device, and has an input means may be used. Typically, a computer having an arbitrary architecture and having a graphic function and a 3D rendering function is preferable. A video game machine, a personal computer, or a mobile phone terminal equipped with a screen is also included. It may also be executed via a network.
[0017]
The calculation means is a means capable of calculation by an arbitrary processor, and is a calculation means selected from an MPU, CPU, graphic chip, DSP, and the like. Further, it may be processed in a distributed manner by a plurality of arithmetic means connected to each other by an appropriate bus or network.
[0018]
Also, as input means, input means that have been widely used so far can naturally be used, such as character value input means such as a keyboard and a voice recognition device, mouse, trackball, pressure sensitive tablet, electrostatic sensitive tablet, etc. This means an arbitrary interface such as a pointing device or a motion capture device that is provided with input data that can be transferred to the arithmetic device as numerical data. Further, the present invention does not limit the operating system. An operating system that appropriately operates the appropriate hardware environment described above is preferable. The method and program of the present invention can be appropriately implemented in any of a multitask OS, a single task OS, an apparatus installation OS, a network OS, and the like.
[0019]
Further, this method and program can be executed in hardware provided with printing means or the like as necessary. Thereby, for example, a comic artist can save labor for comic production work that has been realized manually.
[0020]
“Drawing” refers to drawing a video or image that can be viewed on any display device, storing it in a video memory that outputs it, or printing it as output when the output is obtained on paper. And storage operation in the memory for that purpose.
[0021]
In addition, the expression “continuous” added to gradations and shadows is used in a discrete state quantized to such an extent that it can be viewed as being substantially continuous visually (for example, it is normally regarded as continuous gradation in the field of CG). 256 gradations, 64 gradations, or a gradation display with a small number of colors in which a delicate color tone or shadow is expressed by a lookup table method).
[0022]
DETAILED DESCRIPTION OF THE INVENTION
FIG. 6 and FIG. 7 show examples of images in the middle of generation of comically expressed images realized by the method of the present invention. The image generation flow according to the present invention is shown in FIGS.
FIG. 6A and FIG. 7A show an object obtained by modeling in a virtual space as a preparation stage of the present invention or an object called as data by a contour and a main line. Such an object in a three-dimensional virtual space in which vertical, horizontal, and height are defined is only numerical data expressed as an abstract model in that space. Therefore, rearrangement and various other degrees of freedom are possible, but on the other hand, it cannot be expressed on a two-dimensional display screen surface or paper surface as it is (here, the contour and main line are already represented according to the algorithm described later). It is shown for illustrative purposes only) 6A shows a spherical surface, and FIG. 7A shows a polyhedron in which a part of a cube is dented. Various objects can be used as the object. For example, not only a character such as a person or a car but also a background such as a building or a landscape may be used. Any object that can be defined in the virtual space can be selected as the object of the cartoon expression processing of the present invention.
[0023]
In the preparatory stage of the present invention, the object to be expressed by these cartoon expressions is modeled as a three-dimensional object, or pre-modeled data is read from a file (block 11). Such objects are examined for placement in a virtual space whose coordinates are determined by (X, Y, Z), and their positions, directions, etc. are adjusted (blocks 12 and 13) to determine how to place them. (Block 14). In this way, the arrangement of the three-dimensional object in the virtual space is input (placement input step).
[0024]
Further, parameters such as the viewpoint position and the angle of view (frame) are determined in the virtual space (block 21). In this way, an arrangement to be two-dimensionalized later is determined. In this arrangement, an arbitrary user interface such as a mouse operation is used as used in a normal CG method. In addition, the characteristics of the illumination light source used when rendering later is determined (light source setting step).
[0025]
Further, the object whose arrangement is determined is rendered (drawn) in the two-dimensional image frame 100 whose coordinates are determined by (x, y) (block 22). Although not shown, the rendering step includes a conversion step for perspective-transforming a three-dimensional object in the virtual space onto the virtual screen surface, and a pixel value calculation step for obtaining the color of the object on the virtual screen surface as a pixel value. The actually drawn two-dimensional image frame refers to a memory that holds an actual image of the virtual screen surface and its data.
In the conversion step, a viewpoint provided in the virtual space and a plane (virtual screen plane) provided between the objects are assumed using an object whose arrangement is determined. Thereby, the shape of the object on the two-dimensional image is obtained by perspective-transforming (projecting) the object on the virtual screen toward the viewpoint.
[0026]
In the pixel value calculation step, the luminance on the virtual screen surface indicated by the object surface in accordance with the illumination light is calculated by a calculation method represented by ray tracing or the like, as in the normal method. As a result, a shadow or the like is obtained in consideration of the position and object of illumination in the virtual space, and the position of other objects as necessary. This calculation is performed according to equation (1).
[Expression 1]

[0027]
The first term on the right side of Equation (1) is a term that expresses how the luminance of the surface depends on the inclination between the light source and the surface according to the Lambert rule. In addition, the second term on the upper right side of Equation (1) and the lower right side indicate the illuminance L of ambient light when there is no direct light from the light source. _amb And the coefficient M for the surrounding light _amb This term expresses the effect of. The Lambert rule is a vector L by illumination at the position of the surface (a vector that is directed from the position of the surface toward the light source and the length of the illuminance that the light flux of the light source gives to the surface) and the outside of the object surface. The inner product (N · L) of the surface normal vector N and the diffusion coefficient M of the object surface _dif Multiplied by. The diffusion coefficient is a coefficient determined by the surface material, color, and the like, and includes information on reflectance and gloss. The relationship between the color of the light source and the color of the surface can be considered separately. For example, when the light source is red light and blue light, each light source gives the same illuminance to the surface of the object from the same direction. The reflected light is higher in luminance than the reflected light for blue light. When these effects of the light source and the surface color are taken into account, this expression is evaluated more precisely for each of the three primary colors of R, G, and B. It is not impossible to further reflect the accuracy of photometry and colorimetry. As a simpler method, the shade of the surface may be obtained by glow shading, phone shading, or the like.
[0028]
In the rendering of block 22, the pixel value C for each of the three primary colors R, G, B at all (x, y) coordinates of the virtual screen surface. _i (X, y) (where i is R, G, B) is calculated. Pixel value C _i Is typically from 0 to 255 (when taking an 8-bit value), and the bit width is increased or decreased as necessary. At the same time, a mask M (x, y) indicating whether the object is drawn or not is calculated. The mask (mask information) M is calculated according to equation (2). The mask M is data that becomes 1 if the surface is rendered for the pixel and becomes 0 if the surface is not rendered, and is used when the contour of the object is drawn later.
[Expression 2]

[0029]
The device that executes the block 22 includes an arbitrary storage device that functions as a pixel value memory corresponding to the two-dimensional image frame, and C that is data rendered by the arithmetic device to this storage device. _i (X, y) are stored sequentially. Similarly, M (x, y) is stored in a mask memory having an address space for all pixels.
[0030]
Thereafter, the main line and contour are obtained by a main line / contour drawing step (block 23), which will be described in detail later (block 23). Through this processing, two-dimensional image data represented by main lines (including contours) as shown in FIGS. 6A and 7A is obtained from an object that is a numerical expression in the virtual space. A line 61 in FIG. 6A is a contour line generated from a sphere object, a line 71 in FIG. 7A is a contour line generated from a polyhedral object, and a line 76 is a main line. Such main line / contour drawing is the same even if the object is more complicated.
[0031]
Further, a continuous tone shadow is expressed by one or more textures by a shadow processing step (visual gradation step, block 24) described in detail later. The texture is, for example, a dot pattern shown in FIG. 6B or FIG. 6C or a stripe that is not shown in the conventional screen tone, or “solid” having a certain hue and brightness. For example, “paint”. Accordingly, as shown in FIG. 6B and FIG. 6C, the shadow is expressed by at least one texture having a certain average brightness.
[0032]
Images in which main lines and outlines are created by the blocks 23 and 24 (FIGS. 6A and 7A), and images in which shadows are expressed by textures (FIGS. 6B, 6C, and 7) (B) and (c)) are combined in the image display step (block 25) to obtain an image including a cartoon expression as a whole. Examples of this are shown in FIGS. 6 (d) and 7 (d).
[0033]
As can be seen from FIGS. 6A and 7A, only the main line and the outline are shown, and FIGS. 6B and 6C are used without using the main line. 7 or a combination of FIG. 7 (b) and FIG. 7 (c) only shows the brightness of the surface as a texture. It is difficult to grasp. Although not shown, even if, for example, FIG. 6 (b) and FIG. 6 (c) are simply combined with an image having a continuous real shadow immediately after rendering, a screen tone is added to the real image. It's just like a stretch, far from comical expression.
[0034]
[Main line / Outline drawing]
The step (block 23) for drawing the main line (including the contour line), which is a feature of the present invention, will be described in detail. This main line / contour drawing step will be described with reference to FIG. In this step, first, a target pixel in a two-dimensional image frame is designated (block 30), and a normal vector of an object surface is extracted (block 31). For example, when a three-dimensional object is represented by a polygon, on the surface 5 (FIG. 5) having at least three vertices A, B, and C representing the outer surface of the object, the normal vector N is set for each of the vertices. Determined. A point other than the vertex (for example, D in FIG. 5) can be obtained by interpolating the normal vector N from the normal vector values of the vertices A, B, and C using an appropriate mathematical method. As a result, a normal vector is determined for an arbitrary point on an arbitrary surface of the object. This normal vector corresponds to each direction of X, Y, Z in the virtual space (N _X , N _Y , N _Z ). The normal vector needs to be extracted only for the finally visible surface, and it is not necessary to calculate the surface that is hidden from the viewpoint and is hidden. In this way, the normal vector N of the outer surface at that position can be assigned to the position (x, y) in the two-dimensional image frame. In other words, j is X, Y, Z and N _j (X, y) can be obtained. This is stored in an appropriate storage device such as a normal vector memory. When the surface of the object is determined by another method, the normal vector may be obtained by another method as appropriate.
[0035]
In block 32, depth information (distance information) is extracted. The depth information is the distance from the virtual screen in the virtual space to the outer shape of the object displayed on the pixel, or the distance from the viewpoint to the outer shape. It is determined along the line of sight connecting the viewpoint and the object. If there is a z-buffer value used in hidden surface processing or the like in rendering processing, that value may be used. Unlike the normal vector, it is a scalar quantity, but is obtained as z (x, y) for each pixel (x, y) in the two-dimensional image frame. This may be stored in the depth information memory, or the value of the z buffer may be used as it is.
[0036]
As described above, for each pixel (x, y) in the two-dimensional image frame, the mask M and the pixel value C _i , Depth z, normal vector components N _j Is obtained.
[0037]
The main line is drawn at the boundary between two adjacent planes in the two-dimensional image frame when they are discontinuous in the virtual space and when they are continuous but not smooth. It is also drawn on the contour portion. Although the surface 78 and the surface 79 in the polyhedral object 70 in FIG. 7A are adjacent to each other in the two-dimensional image frame 100, they are discontinuous in the virtual space. In addition, the surface 77 and the surface 79 are adjacent to each other in the two-dimensional image frame and are continuous in the virtual space, but are bent and are not smoothly continuous. In comics and the like, the boundary of such a surface has been drawn by the creator by a line called a main line.
[0038]
To draw this main line, for each pixel (x, y) in the two-dimensional image frame, the neighborhood of that point and z or N _j The difference of whether or not has changed is obtained (blocks 33 and 34). These can be performed by a Laplacian filter, and for z, Equation (3), N _j For, a difference can be obtained by equation (4).
[Equation 3]

[Expression 4]

[0039]
By using this difference value, the presence / absence of a main line and its line width are obtained. For example, W obtained from z by equation (5) _z Is obtained from N by Equation (6). _N Are obtained, and the line width is obtained by using an appropriate function for those values.
[Equation 5]

[Formula 6]

[0040]
Here, [constant z] and [constant N] are offset values indicating that a main line is not drawn when a difference value greater than a certain value cannot be obtained. W _z , W _N When is negative, no line is drawn. In block 35, the mask information M (x, y) that has already been calculated by the equation (2) is extracted.
[0041]
Also, the contour line is obtained. Since the mask M has been obtained, the contour can be determined using this mask (block 36). That is, the sum of the values of the mask information M of 9 pixels including the target pixel and the surrounding 8 pixels is calculated, and when it is 0 or 9, the pixel is inside or outside the drawing area of the object. In this case, it is indicated that at least one of the target pixel and the surrounding pixels includes the boundary of the projected figure on the two-dimensional image frame of the object. Therefore, it is possible to determine that the target pixel is a contour (block 36). Here, a variable called contour coefficient B is used. The contour coefficient B is a value larger than 1 (a value that can be set by the operator) when it is determined that the pixel of interest is a contour, and is 1 when the pixel is not a contour. As a result, the formula for giving the line width of the main line and the formula for giving the line width of the contour line can be defined together, and the line width W (x, y) becomes as shown in formula (7).
[Expression 7]

[0042]
Using the line width W obtained in this way, a circle having a radius corresponding to the line width W is placed at a position (x, y) on an appropriate layer (hereinafter referred to as “main line layer”) superimposed in the two-dimensional image frame. ) At the center (block 37). Scanning the pixel of interest within the two-dimensional image frame results in a main line and contour. The main line layer is stored in an appropriate storage device such as a main line memory. Finally, the main line layer data is displayed in a two-dimensional image frame. The main line layer can be appropriately changed in color, and can be drawn in black or other suitable colors. In the case of separately printing comics or the like, the color of this portion may be selected at the printing stage.
[0043]
After finishing the processing for a certain target pixel in this way, the target pixel is moved (block 38), and scanning is completed over all the pixels (block 39). When moving the target pixel to the adjacent pixel, for example, if the target pixel is moved immediately to the right, the number of accesses to the memory for depth and normal vectors is reduced. Only the right three pixels can be read from the memory, and the remaining values of the target pixel and the surrounding pixels can use the values of the target pixel at the immediately preceding position.
[0044]
In addition to the order described in the above description, for example, it is possible to process the block 33 immediately after the block 31 and then process with the block 32 and the block 34. Mask M necessary for drawing main line / outline of block 37, difference value dz of depth, difference value dN of each component of normal vector _j As long as each value of is obtained for the pixel of interest.
6A and 7A are obtained by the main line / contour drawing step as described above. In FIG. 7A, both the contour and the main line are drawn, but the contour portion is drawn thicker than the main line.
[0045]
Although not shown, depending on the type of object, the main line can be drawn in advance by a technique such as texture mapping on the boundary line (surface border) that surrounds and forms the individual surfaces that make up the object. . This is because when the object is composed of a relatively small number of faces and is a combination of straight polyhedrons (such as buildings), the main line is represented by the existing texture mapping method at the rendering stage. It is valid. In this case, the main line can be clearly shown in the two-dimensional image frame, but the main line and the contour line cannot be determined, and the curved surface may have a contour other than the boundary line. In this case, the contour can be expressed by a main line by using the contour drawing step together.
[0046]
[Shading processing (visual gradation step)]
Further, another shading process step (visual gradation step, block 24), which is another feature of the present invention, will be described in detail with reference to FIG. In the conventional cartoon expression, shadows appearing in a picture are not necessarily expressed with continuous tone, but are expressed with one or more textures. As the texture, a pattern called a screen tone having a binary contrast such as periodic black and white is often used in many cases. The screen tone may be, for example, a square lattice or a triangular lattice of circular dots (FIG. 6B and FIG. 7B are drawn using a similar pattern). These are made by arranging a large number of colored circles such as black having a constant pitch determined by the number of screen lines (number of lines per inch = 2.54 cm) on a transparent sheet. In many cases, a pattern having about 60 to 70 screen lines is used. In addition, stripes and various other screen tones are used for cartoon expression. In addition, uniform solid coating or the like can be used as the texture. The shadow processing step (visual gradation step) of the present invention is not limited, but such texture is reproduced on a computer, and the shadow that the object is realistically expressed is used as a comic like it. This is a process for expressing the expression texture.
[0047]
In this shading process, first, a pixel of interest is designated (block 40), and a gray scale process (block 41) is performed as necessary. This is an equation for calculating the Y signal of the video signal at each point of the two-dimensional image frame, for example, according to Equation (8).
[Equation 8]

[0048]
If necessary, the brightness and gradation of the image are adjusted by performing gamma processing or the like (not shown). And the C of that pixel _Y Reference value T that (x, y) is predetermined by an operator or has a default value ₁ , T ₂ Compare with etc. The reference value defines the number of layers (hereinafter referred to as tone layers) that store textures that express shadows, and is designated by the number obtained by removing 1 from the number.
[0049]
Next, each reference value T _k (Where k = 1, 2,... (Number of tone layers) -1) and C _Y And compare. As a result, for example, the number of tone layers is 3, and the reference value is T ₁ <T ₂ If only two values are defined, T ₁ , T ₂ A tone layer to be drawn at the position (x, y) is determined with reference to. For example, C _Y <T ₁ (Block 42), the pixel of interest is drawn with the texture stored in tone layer 1 (block 43). ₁ <C _Y <T ₂ (Block 44) with the texture stored in tone layer 2 (block 45), T ₂ <C _Y (Block 46), the texture stored in the tone layer 3 is drawn (block 47). The pixel of interest is then moved (block 48), and all pixels are scanned (block 49) to perform this process in the two-dimensional image frame. As a result, for each of the tone layers 1, 2, and 3, using the density defined by the occupation ratio of the circular pattern in advance, for example, density 50% (dark), density 20% (slightly dark), density 0% ( If a texture such as “nothing” is stored, it is possible to express a shadow as seen in a cartoon expression in which each gradation is expressed by a texture such as a screen tone after the gradation is reduced.
[0050]
FIG. 6B and FIG. 7B both use the expression of a screen tone having a density of 20%, and are the data written in the portion to draw the data imitating the screen tone in the tone layer 2. FIG. 6C and FIG. 7C are similar data drawn from the tone layer 1 using a screen tone expression of 50% density. The expression of the screen tone having a density of 0% is the same as when nothing is drawn, and is not particularly shown. The texture may be prepared in advance and prepared in an appropriate storage device, but may be generated appropriately using some mathematical relationship and numerically generated. For example, C divided by judgment value _Y A dot having the same density can be generated at the same value. Alternatively, the original pixel value C _i The texture may be generated by processing such as maintaining the hue and reducing the brightness.
[0051]
If the main line layer prepared in block 23 and at least one or more tone layers prepared in block 24 are combined and displayed as an image (block 25), FIG. 6 (d) and FIG. 7 (d). Is obtained. As a result, the operator can confirm a two-dimensional image on which a cartoon expression generated from the three-dimensional object data is applied as an image.
[0052]
In an application in which the purpose is achieved by image display (for example, a computer game), the image is processed by blocks 22 to 25 while the object is driven by a geometry engine as necessary. The frame is advanced to generate a continuous movie.
[0053]
In addition, at an animation production site or the like, it is determined in block 26 whether or not the request is made by the operator as necessary. As a result, the blocks 22 to 26 are re-executed from the appropriate points A, B, and C, and the processing operation of the CG including the cartoon expression is efficiently performed on the computer.
[0054]
In addition, in applications that are not provided to the user in the end but are provided by printing (DTP, comic production site, etc.), further, including a printing step (block 27), depending on the result Blocks 22 through 27 are re-executed from the appropriate A, B, C point.
[0055]
In the method of the present invention described above, the main line / contour drawing step (block 23) and the shadow processing step (block 24) do not have to be performed in this order. Conversely, the main line / contour drawing step may be performed after the shading processing step. Since most main lines and contour lines are overwritten by the shadow texture, they can be directly written into a two-dimensional image frame memory (frame buffer) for image display. In addition, the pixel of interest is scanned independently inside the block 23 and the block 24, but after adjusting the use of pixel data around the pixel of interest in the difference processing, the scanning loop is made common. Also good. In this way, the difference value of z, the normal vector component N _j It is possible to save the memory for storing the difference values of the two-dimensional image frames.
[0056]
[Example]
As shown in FIGS. 8 to 11, a cartoon expression was constructed from building objects. FIG. 8 shows the arrangement of this building in the two-dimensional image frame. As in FIGS. 6A and 7A, the object itself in the virtual space is numerical data, but for the sake of explanation, the figure after the main line / contour processing of the present invention is shown.
The object building 80 is arranged in a virtual space according to the method of the present invention. The object also includes a portion 81 that can be seen on the ground.
[0057]
Further, the operator determines the illumination to be applied to the object 80, and determines the viewpoint and angle of view for observing the object in the virtual space. In this embodiment, a parallel light source placed almost directly above is set.
Rendering (conversion step, pixel value calculation step) was performed on the building object with this light source. Immediately after rendering (immediately after the processing of block 22 in FIG. 2), although not shown, the brightness varies greatly from surface to surface, and the brightness varies depending on the location within the surface (this is the roof in FIG. 11). Are shown as a bright portion 87 and a dark portion 86). In the stage immediately after rendering, this difference in brightness is felt as a continuous shadow (gradation), which is a so-called realistic expression, but is not an image expression seen in comics or the like.
[0058]
After this rendering, the main lines and contours of the present invention were drawn, and shading processing (visual gradation) was performed. The main line / contour drawing was automatically generated from the distance information between the surfaces of the solid object and the normal vector direction according to the method of the present application. If you change the parameters (coefficients, constants) of the functions in Equation (5) and Equation (6), you can change the line thickness, choose to pick up all the borders to be drawn as main lines, and avoid picking up too much It was possible to choose. In FIG. 8, a line 82 represents a contour line obtained in the present embodiment, and a line 83 is a main line. The contour line coefficient B was set to be 1.5 times when the main line is a contour.
[0059]
In addition, the shadow was expressed with three types of textures including blanks. FIG. 9 shows an arrangement 84 of a 30% halftone dot pattern assigned to the intermediate lightness portion in this embodiment. The texture stored in the layer 2 is cut out to a necessary range. FIG. 10 shows an arrangement 85 of a halftone dot pattern assigned to the dark lightness portion, in which the texture stored in the layer 1 is cut out to a necessary range. Layer 3 is blank here. This assignment is made up of C values obtained from RGB pixel values. _Y T is set so that the value of (Y value) is 0% to 31 with 50% halftone dots, 32 to 127 with 30% halftone dots, and 128 to 255 with blanks. ₁ 32, T ₂ Is obtained as 128. Each pixel value is obtained by rendering based on the diffusion coefficient (including reflectance and gloss information) determined for each object attribute, the lighting conditions, the relationship between the direction of lighting and the surface, the shadow created by the object, etc. Pixel value. 9 and 10 are both automatically generated from the three-dimensional object data by performing the shading process (visual gradation) of the present invention.
[0060]
The main line / contour image generated as described above and the texture image are combined to generate a comically represented two-dimensional image shown in FIG. 11 and displayed on the display device. The part 87 and the part 86 in FIG. 11 are both the same surface of the same object, but are drawn with different textures reflecting the gradation seen in this part immediately after rendering. The main line can be easily drawn by changing the parameters of the equations (5) and (6) to determine how much detail is drawn as the main line. Moreover, by changing the contour coefficient, the thickness of the main line (contour line) drawn on the contour can be coordinated easily.
[0061]
Also, a reference value T used as a reference value for adopting the texture ₁ , T ₂ As a result of the operator's adjustment, the portion to be replaced with the halftone texture has changed, and the processing state can be easily adjusted. The textures are prepared in addition to the halftone dots, and various shades with different impressions can be easily confirmed by switching between them.
[0062]
Further, the operator can easily change the orientation and arrangement of the object in the virtual space, and can easily change the viewpoint. As a result, the objects can be rearranged and the parsing process (feeling of depth) can be easily corrected. In addition, the same object could be used in a different direction and reused.
As described above, due to the features of the present application, a comic-processed two-dimensional image intended by the operator can be easily generated from the three-dimensional object data.
[0063]
【The invention's effect】
According to the present invention, in computer graphics technology, a main line or outline conventionally used in comics or the like is drawn, and a texture such as an arbitrary pattern or a solid color of a fixed gradation is used for a shadow. Expression becomes possible. This allows conventional expressions to be used in computer graphics. In addition, the advantage of a computer that allows objects to be rearranged easily can be used in comic production and animation production sites.
[Brief description of the drawings]
FIG. 1 is a flowchart showing a method for arranging a three-dimensional object according to the present invention.
FIG. 2 is a flowchart illustrating a method for generating a two-dimensional image from stereoscopic object data according to the present invention.
FIG. 3 is a flowchart illustrating details of a main line / contour drawing step of the present invention.
FIG. 4 is a flowchart illustrating details of a shadow processing step according to the present invention.
FIG. 5 is a perspective view showing the concept of normal vector calculation according to the present invention.
FIG. 6 is an explanatory diagram showing a manner in which a two-dimensional image of a cartoon-represented sphere is generated from a sphere object by the method of the present invention.
FIG. 7 is an explanatory diagram showing a manner in which a two-dimensional image of a cartoon-representing sphere is generated from a polyhedral object by the method of the present invention.
FIG. 8 is an explanatory diagram showing a two-dimensional image expressed from a main line and an outline generated from a building object according to the embodiment of the present invention.
FIG. 9 is an explanatory diagram showing a texture (20% halftone dot) pattern generated from a building object and a portion represented by the pattern, according to an embodiment of the present invention.
FIG. 10 is an explanatory diagram showing a texture (50% halftone dot) pattern generated from a building object and a portion represented by the pattern, according to an embodiment of the present invention.
FIG. 11 is a diagram illustrating a cartoon-represented building two-dimensional image generated by combining a main line and contour generated from a building object and a texture pattern according to an embodiment of the present invention. FIG.
[Explanation of symbols]
23 Main line / contour drawing
24 Shading process (visual gradation step)
100 2D image frame
60 ball object
61 Main line (contour)
70 Polyhedral objects
80 Building objects
601, 602, 701, 702, 84, 85 Texture (dot pattern)

Claims (5)

A method for generating a two-dimensional image of a cartoon expression from three-dimensional object data,
An arrangement input step for accepting an arrangement of a three-dimensional object in a virtual space by an input means;
A viewpoint setting step of accepting the position of the viewpoint in the virtual space by an input means;
A light source setting step of accepting the characteristics of the illumination light source in the virtual space by an input means;
And obtaining by calculating means as the color information about the color indicated by the solid object by the illumination light source, each of the display pixels,
Based on the color information and a certain reference value, by assigning a texture having a different brightness to any of the two-dimensional image portions divided by the reference value in the two-dimensional image, each display pixel is assigned. On the other hand, a visual gradation step in which the computing means adopts and draws texture pixels having different brightness before and after the reference value;
Obtaining the position of the display pixel of the two-dimensional image where the three-dimensional object is present from the viewpoint as mask information by a calculation means;
Obtaining a distance between an external surface that can be seen from the viewpoint of the three-dimensional object and the viewpoint by calculating means as distance information for each of the display pixels ;
Obtaining a normal vector in the virtual space of the outer surface by calculating means as normal vector information for each of the display pixels ;
After the visual gradation step , based on the mask information, the distance information, and the normal vector information , a line that becomes a boundary between a plurality of surfaces of the stereoscopic object in each of the display pixels or the stereoscopic object A main line drawing step for drawing data representing a main line for emphasizing the outline of the line by an arithmetic means;
The method comprising an image display step of displaying the image drawn data of the main line in the image of the texture adopted in the two-dimensional plane by the computing means. The method for generating a two-dimensional image of a cartoon expression from the three-dimensional object data according to claim 1, wherein in the main line drawing step, a main line indicating the outline of the three-dimensional object has a line width different from a main line indicating other than the outline. The main line drawing step includes:
Obtaining a difference value of the distance information from the distance information between the target pixel of the two-dimensional surface and the surrounding pixels;
Obtaining a difference value of the normal vector information from the normal vector information between the target pixel of the two-dimensional surface and the surrounding pixels by a calculation means,
2. The cartoon expression from the three-dimensional object data according to claim 1, wherein the main line drawing step determines a main line drawing by a calculation unit based on the difference value of the distance information and the difference value of the normal vector information. A method for generating a two-dimensional image. 2. The three-dimensional object data according to claim 1, wherein the visual gradation step performs gradation using a numerical value generation texture generated by calculation means using a numerical value calculated from the color information and the reference value. To generate a two-dimensional image of a cartoon expression from an image. The program for implement | achieving the method in any one of Claim 1 to 4.

Images